OpenAI announced a proof addressing the Navier-Stokes problem, one of the Clay Mathematics Institute's Millennium Prize challenges, generated with help from an internal large language model. Mathematicians, including Luis Silvestre of the University of Chicago, argue the proof relies on introducing an external force that alters the equations into a version disconnected from real fluid behavior. Three mathematicians have since posted their own proof showing OpenAI's method cannot be extended to resolve the original, unmodified problem.
scientificamerican.com
· 2026-09-22
Spain's Data Protection Agency (AEPD) has received its first notification of a breach allegedly executed by an autonomous AI agent built on a large language model. According to the report, the agent found system vulnerabilities, logged in, probed connected applications, then altered personal data and accessed financial records. The AEPD has not yet verified the claims but says the case demonstrates that AI-driven breaches have moved from theory to practice.
bleepingcomputer.com
· 2026-09-16
Satirical creator Graham Zip tested an AI-powered job interview avatar named 'Dana Whitfield' by responding with strings of random, meaningless words instead of real answers. The chatbot failed to recognize the nonsense, instead confidently paraphrasing and affirming the gibberish as if it were coherent professional experience.
futurism.com
· 2026-09-05
Chinese AI firm Z.ai's GLM-5.3 model helped find over 1,000 critical vulnerabilities in open-source projects including the Linux kernel, adding to a wave of AI-discovered bugs this year. Kernel maintainer Greg Kroah-Hartman showed data indicating CVEs fixed per kernel release have surged from about 500 to nearly 2,000 as AI code-review tools improve.
techspot.com
· 2026-09-02
An Entrepreneur contributor argues most companies claiming to be 'AI companies' have simply bought licenses and run pilots, not restructured how they operate. The piece defines true AI-native organizations as ones built on the assumption that intelligence is cheap and abundant, where workflows, staffing and measurement would collapse without the models—unlike adopters, who could remove AI tools and keep functioning, just slower.
entrepreneur.com
· 2026-08-25
Thomson Reuters unveiled Thomson, its first proprietary large language model, which it built by fine-tuning an open-source base rather than training from scratch. The company says it spent roughly $40 million on the project, using its own legal, tax and news archives from Westlaw, Practical Law, Checkpoint and Reuters alongside hundreds of subject-matter experts to specialize the model for professional work.
thomsonreuters.com
· 2026-08-25
Pangram, a commercial AI-text detector built by New York start-up Pangram Labs, is being adopted by scientific publishers, computer-science conferences and universities to flag undisclosed AI-generated content. The American Association for Cancer Research uses it to catch peer reviewers who submit AI-written reports without disclosure, while NeurIPS rejected 18% of paper submissions after running them through the tool, and arXiv preprints can now be checked via a mirror site called alphaXiv.
nature.com
· 2026-08-25
OCR It is a Chrome extension that lets users draw a fixed capture region over a paginated document viewer—like a scanned book or locked PDF—and repeatedly screenshot and OCR that same spot with a hotkey, building a full text transcript as pages turn. An auto-run mode can capture, advance the page, and repeat until the document ends, all processed locally using a bundled Tesseract engine with no network calls or API keys required.
github.com
· 2026-08-24
GoKawiil